5 ms·

There's a great blog post[1] by one of the OpenBSD developers about why they did so. tl;dr using bitmasks necessitates namespaced enums/defines that take up hor

There's a great blog post[1] by one of the OpenBSD developers about why they did so. tl;dr using bitmasks necessitates namespaced enums/defines that take up horizontal space, strings are easier and don't need to go through the C pre-processor. [1] https://flak.tedunangst.com/post/string-interfaces https://flak.tedunangst.com/post/string-interfaces

agileAlligator 4y ago
great find!
hag 4y ago
Nice! I would probably have used bitmasks in this situation, but as usual there is a reason behind the choice and I get the reasoning.
IshKebab 4y ago
Nice find. That article is highly unconvincing though and mostly argues against straw men. > Although using strings subverts C’s already weak type checking, that’s probably not a major concern. One can screw up bit masks by using || in place of |. Or, as above, one can incorrectly pack the magic array. It’s usually much easier to visually audit a string than the C code used to plaster a dozen option together. It's pretty easy to design an interface that is way way less error-prone than strings (especially ones full of single-letter differences!) and the visual auditing argument falls apart as soon as you have to `snprintf()` some string together from parts. This code is way more readable, way less error prone, more discoverable, faster and more easily extendable than strings: auto config = make_pledge_config(); config.read_path = true; config.stdio = true; pledge(&config); You'd think security focused people would care about static type checking.
- jeshin 4y ago
  if you really care about static type checking, you probably wouldn't be using C
  - IshKebab 4y ago
    You probably would if you care about static typing and are working on a kernel syscall interface.
    - jeshin 4y ago
      and yet they chose to use strings for their api, eh?
      - IshKebab 4y ago
        Well exactly. That's why it's so weird.

agileAlligator 4y ago

great find!

hag 4y ago

Nice! I would probably have used bitmasks in this situation, but as usual there is a reason behind the choice and I get the reasoning.

IshKebab 4y ago

Nice find. That article is highly unconvincing though and mostly argues against straw men. > Although using strings subverts C’s already weak type checking, that’s probably not a major concern. One can screw up bit masks by using || in place of |. Or, as above, one can incorrectly pack the magic array. It’s usually much easier to visually audit a string than the C code used to plaster a dozen option together. It's pretty easy to design an interface that is way way less error-prone than strings (especially ones full of single-letter differences!) and the visual auditing argument falls apart as soon as you have to `snprintf()` some string together from parts. This code is way more readable, way less error prone, more discoverable, faster and more easily extendable than strings: auto config = make_pledge_config(); config.read_path = true; config.stdio = true; pledge(&config); You'd think security focused people would care about static type checking.

jeshin 4y ago
if you really care about static type checking, you probably wouldn't be using C
- IshKebab 4y ago
  You probably would if you care about static typing and are working on a kernel syscall interface.
  - jeshin 4y ago
    and yet they chose to use strings for their api, eh?
    - IshKebab 4y ago
      Well exactly. That's why it's so weird.